Software Effort Estimation using Regularized Radial Basis Function Neural Networks

摘要

The value of Artificial Neural Networks (ANNs) methods in performing complicated pattern recognition and nonlinear estimation tasks has been demonstrated across an impressive spectrum of applications. ANNs methods has been used extensively, in the software cost estimation process, due to the complexity of the relations between the project's attributes. ANNs Radial Basis Function (RBF) networks have advantages of easy design, and strong tolerance to input Noise. This paper, studies the effect of using Regularized Radial Basis Function Networks in improving the accuracy of the software cost estimation, using different training algorithms like K-means method, Genetic algorithm and Particle Swarm Optimizer (PSO). The relative improvements were found to be around 40 %, by using the Regularized Radial basis function with the PSO Algorithm.